Larian Studios Forums Baldur's Gate III Baldur's Gate III - Suggestions & Feedback Here we are again....with another Loaded Die rant

I have been acutely observing the attack rolls of both my team and enemies in the combat log.

And I have came to a conclusion (my Karmic Dice is off):

1. Enemies have a very high chance to roll 18 and above. Almost as if the game knows that if the enemy rolls were perfectly fair, the player would win easily. So the game gives them a handicap. The best example of this was when i fought Yurgir in Gauntlet of Shar. He had many minions. I once had 6 minions attack, and all rolled 18s, 19s and 20s.

2. Your characters have a very high chance of rolling 4 or less.

This is exactly why 80% feels like the actual 50%. When you have an 80% chance to hit, you miss half of the time. But when your enemies have only a 50% chance, they hit 4 out of 5 times. And I can say with certainty than I get far more Critical Misses than I do Critical Hits. It's due to the skewing of rolls for the player to get 4 or less. So your Critical Miss chance is not 5%....it's more like 15% to 20%.

I don't know what to tell you, man, except suck it up, you have a bad rolls.

I had one too.

I had a day where rolling for a DC 30 lock is nigh immpossible. The day after I rolled 30+ more than 5 times for 10 DC lock.

I shot a lot of point blank range attack 65% odds Disadvantaged and hit it. The mob -rarely- hit me when I stand right in front of them when they try to shot me point blank.

What I'm trying to tell you is that you and many other people have some kind of loss aversion.

I suggest you pen down every single roll number, down to perception/survival/nature checks you get, everything.

Are you using Karmic Dice (it's a game setting that defaults to on)? If so...yes. Karmic Dice is known to even out the odds of rolls such that the target DC (or AC) matters less. Enemies (and players) roll higher values against higher DCs. It also maybe does something to prevent you from succeeding or failing too many times in a row, adjusting the odds based on previous rolls.

Turn Karmic Dice off if you want "pure" rng.

If you've already turned Karmic Dice off, then how many rolls are you basing this on? If less than a few hundred, then it very well might just be rng combined with the human tendency to notice & remember bad things over good things. You need ~500 or more rolls to be able to declare anything about the rng on something with 20 different outcomes.

The dice ARE loaded. Best I can tell you is save scum until Larian do something about it. As per my own research, skill checks seem pretty on point and fair. Attack rolls are another story. I recorded every ATTACK roll in act one without save scuming and averaged them out. I broke it down into AI and PC. Pc rolls being an attack roll from anyone that is currently in my party. I had Karmic Dice off for "the TT experience" roughly 1200 rolls from each side and my findings were the "player rolls" were averaging around 7.4 and enemy rolls are averaging 10.9. So get used to hitting the F5 button before your good attacks and then the F8 button when you inevitably miss it with a stupidly low roll.

Footnote: these averages are the best I could get being when you have advantage or disadvantage it only tells you the die roll that is used. In my play through I had advantage far more than disadvantage so that 7.4 should technically be lower, but being disadvantage take the lower of 2 rolls and I can't see what the higher roll was to properly calculate the averages, I'm going to say they cancel each other out.

I'm guessing you're playing on Tactician?

Which means every enemy roll got +2.

I mean, it's pretty normal for enemies to have advantageous roll, IF the roll indeed loaded, which they're not, it literally needs one person to do the same thing you did and have different result for you to be wrong.

Look, I'm not saying you're wrong, because your dice rolling experience has been bad. I'm saying your method might not exactly correct to convey the thing you want to confirm.

If you roll a nat 1, once and nat 20 twice, your average is around 13,x. If reversed (twice nat 1, a single nat 20), it's 7,x. It doesn't prove anything other than averages are varies depending on the number.

This thread will not be the last nor the first of the similar complain. Pretty sure at least hundreds of thread in any game with D20 roll system revolves around the rolls are rigged.

How does that interact with critical miss and critical hit anyway? Does that mean they never critically miss and have a 15% critical hit chance?

It seems a very strange way to increase difficulty in a D&D game if you could just buff the enemies fundamental properties, like attributes and level, such as giving a +4 to their main stat and a +2 to all other stats.

Doesn't matter if the enemies hot more often, when my tooltip says 50% its more like 5%.

When it says 85% is actually 50%.

95% is actually around 65-75%.

There is something hugely wrong with the success chance displayed in game.

Could you post that data, in a google spreadsheet or something? It's been a pet project of mine to analyze the rng throughout EA, but I don't yet have any large data sets for dice rolls in the Full Release.

1200 rolls is enough to be fairly robust...except for the whole "mix of normal and (dis)advantage rolls." There's probably some statistical method to account for that unknown...

I would But I didn't want to deal with tabbing in and out so I used pen and paper. My wife cleaned the living room and threw them away on me, I honestly thought I was the only one nerdy enough to want to see those numbers. While I'm curious to to it again to see if I get the same results, I'm going to wait until a modder gets frustrated as I did and makes a mod that will automatically export Rolls to a spreadsheet.

If you roll a die the average is going to be half the dies value +1, so average roll for a d6 will be 4, a d8 will be 5, a d10 will be 6, d12 will be 7 ect.
What experience was, somewhere around 1000 rolls the rolls should of been 11ish on both sides, keep in mind these are the unmodified rolls I'm speaking of. so if you have plus 5 to hit and roll a 7 your total would be 12, I'm talking I only recorded the 7. The pluses are something the player has an effect on via stat gear spells what have you. If you want to say true RNG could be anywhere that's fine but the AI "rolling" a value that is 4 higher than the player on the regular is just down right unfair, and in this situation would be cheating.
And no I was not playing tactician mode. I was playing on balanced as I was hoping that would offer up the most authentic TT game play.

I had an amusing experience earlier.

Just started a new "no reload" sort of ironman game after Patch 1 came out. I failed to talk Gimblebock and crew from leaving, so I had to fight them. I got to the door that Andorn was guarding and failed to persuade him that Gimblebock had triggered a trap and needed help.

So I had to try lockpicking the door. I had six sets of lockpicks. And I was using Astarion along with Guidance.

I failed all six rolls. All I could do was shrug my shoulders and head for the grove.

*

I just don't have enough data to know if the rng is off. There are times where it feels like it, but I know I probably have some anecdotal bias. I do hope someone manages to figure out a way to export some rolls so a real analysis can be done.

Hmmm so you recorded 2400 rolls and then manually added them up (in sets of 1200) to get the two averages? That's...a lot of data to do math on, whether you did it by hand or used a calculator. Transcription & math errors might have a significant effect: accidentally entering "155" instead of "15" once will will shift your average by ~0.1. Those uncertainties still *probably* wouldn't explain the 3.5 difference you found between player and enemy, but...

Not quite. It's half the die's value plus 0.5.

I suspect that this is how of works, I will be doing an in depth video tomorrow and post it here. I will be testing dice rolls as I have had enough and I want to set the record straight. Now as to my current theory....

The odds of hitting are displayed as 60%, 35%, 95%, as these are numbers that are bogus, if you tell a player 60% when their odds are 40% it will be less frustrating to the player, and give them more hope of hitting, while of they have 40% [actual] they may be discouraged and quit playing. As anyone who understands math will tell you, when you roll a dice, there is NO WAY to get any odd numbers. It's always 10-100%. If enemy has 16 AC and your hit is 14, the odds of hitting is 0%. I taken stats where I know my odds of hitting an enemy should be 0%, still showed it as 20%

Likewise, the %to hit is before the odds of critical or critical miss. So if your odds of hitting is 65%. The odds are 12/20, but in actuality are 11/20. So I believe is blatant lying to us.

10.9 isn't too bad (vs 10.5, that sounds acceptable), but 7.4 is despicable.

I will never forget the days of EA when my paladin had ungodly luck and my druid kept failing even the nature checks. At some point, I had the paladin do them. He passed.